Learning and understanding single image depth estimation in the wild

Talks & Slides

Introduction

A gentle introduction to single image depth estimation[TALK] [SLIDES]

Stereo supervision

How to train a single image depth estimation network with stereo images, state-of-the-art and related problems[TALK] [SLIDES]

Mono supervision

How to train a single image depth estimation network with monocular sequences, state-of-the-art and related problems[TALK] [SLIDES]

Understanding single image depth estimation

How neural networks learn to estimate depth from a single image and how reliable estimated depth maps are[TALK] [SLIDES]

Auxiliary supervision

How to obtain stronger supervision from various imagery and additional tasks[TALK] [SLIDES]

Learning single image depth estimation in the wild

How to combine several collections of images to obtain a robust network ready for in-the-wild deployment[TALK] [SLIDES]

Mobile depth estimation

How to design a real-time, robust monocular network running on a smartphone (Apple iPhone XS)[TALK] [SLIDES]

Conclusion and discussion

Closing remarks and discussion about current limitations and future challenges in this field[TALK] [SLIDES]